Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

A different discriminator? #10

Open
ehsan-soe opened this issue Feb 2, 2020 · 5 comments
Open

A different discriminator? #10

ehsan-soe opened this issue Feb 2, 2020 · 5 comments

Comments

@ehsan-soe
Copy link

ehsan-soe commented Feb 2, 2020

Hi,

I really enjoyed your paper.
In this regard, I have following questions that I appreciate your reply,

  1. I wonder if it possible to replace the current discriminator which is build on top of the LMHead with any other trained discriminator? In other words, does it really need to be built on top of the LM (p(x)) itself, or it can be any discriminator?
  2. If the answer to previous is yes, then is it possible to fine-tune gpt-2 on our own data and then generate from that with our specific discriminator (attr model)?
@mshlis
Copy link

mshlis commented Feb 3, 2020

@ehsan-soe

  1. It needs to be built on top of the LM to get gradients for backprop (if otherwise you would have to use some form of expensive/noisy gradient approximation scheme to use the same approach)
  2. You can still fine tune this approach on your own data, just fine tune the descriminator ontop of the LM

(Note: i have no actual association to this project)

@srulikbd
Copy link

@mshlis so can I just replace GPT2 with like "distilbert-base-multilingual-cased"?

@mshlis
Copy link

mshlis commented May 10, 2020

@srulikbd sure but for what result? The goal of this is to do text generation which you can’t do effectively with distilbert. You probably want a backbone that is trained on next word/letter prediction.

@srulikbd
Copy link

OK, you are right.
I saw experiments that tried to use BERT like LM but got bad results, because of the bidirectionality training.
so I think I should find GPT2 that was trained on my specific language, or train one myself..

thanks for the quick answer!

@liya-gafurova
Copy link

liya-gafurova commented May 20, 2020

Hello! Thank you for your great work!

In your article you have mentioned that instead of GPT2 it could be used another transformer-based text generator ("but the method applies in any representation space from any transformer-based text generator").
Is it possible to replace GPT2 with CTRL? And how to manage using CTRL with its different codes (like "Links", "Wikipedia", "Rewies", "Reddit " etc) within PPLM codebase?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants